jstokes | # 1341 | 2011-03-13, 22:11 | Report

Are you using the Diablo 5 theme? If so, what order did you install them in? Because Diablo 5 comes with a workaround that gets setup when installing the theme. Here's the relevant excerpt from its postinst:
If you're not using a dark theme, just search for "Tear" in the Diablo5 thread as it has the CSS there somewhere anyway

Snoshrk | # 1342 | 2011-03-13, 22:30 | Report

Doh! No wonder my Search-Fu failed...wrong thread.

Yep I am using the Diablo5theme....pretty sure it was installed first...

Will try the work around asap

Thanks!

Kroll | # 1343 | 2011-03-14, 22:29 | Report

If you use Diablo 5 theme it is a bag which happens because tear doesn't use gtk input fields theme. Sorry, but this bag is unfixable. The only way is to use other browser or other theme.

Snoshrk | # 1344 | 2011-03-15, 01:26 | Report

[WHINE] But I LIKE both! [/WHINE]


Oh well I guess I'll just have to use the Zen input method in Tear.

Thanks

mcow | # 1345 | 2011-05-30, 16:10 | Report

Hi all,
I've got the 0.3.5-2 version of Tear installed. Reading this thread, I discovered there's a 0.3.6 but it seemed like there were some additional dependencies - enough hassle that I haven't tried it yet.

I am curious how 0.3.6 scores on the HTML 5 test. 0.3.5 scores 95, which isn't horrible for this form factor.

What I'm really after is HTML5 EventSource support, which is detailed near the bottom of the test page under "communication". I have a personal project I'm working on that works best on the N800 in Tear. It would be a lot cleaner, and draw down the battery a lot less on my tablet, if I could use event sources rather than polling. If someone tells me that EventSource is supported in 0.3.6, I'll take the plunge right away.

tso | # 1346 | 2011-05-30, 17:25 | Report

Iirc, both use the same webkit engine. The difference is largely related to the interface.

Mitrandir | # 1347 | 2011-05-30, 19:28 | Report

Yes, support for HTML5 features is all up to WebKit engine. And, as far as i know, Bundyo had a huge problems while trying to update Tear to newest versions of webkit due to damned old GLib. So, chances that Tear will receive support for new html5 features are very low...

Anyway, i don't think that support for EventSource can save some battery or cpu cycles on N8x0. Long-polling techniques (comet) already uses pretty low cpu and traffic. Significant improvements are only possible with carrier support as discussed in whitepaper: http://dev.w3.org/html5/eventsource/#eventsource-push
But for Maemo there is no any carrier-provided push notifications, and TCP-connection (probably with some keep-alives) is still neccessary for receiving events. But this is not much better than long-polling (which also more compatible and cross-browser).

---

Last edited by Mitrandir; 2011-05-30 at 19:31.

mcow | # 1348 | 2011-06-04, 21:52 | Report

Thanks, Mitrandir. I reworked my project to use long-polling and it's using much less CPU on the tablet.

mcow | # 1349 | 2011-06-12, 21:25 | Report

I have a widget in the page of my project which I would like to be able to long-press and/or double-tap. In Tear, I see the following:
1) I'm not getting the doubleclick event when I double-tap. Is this operator error or is double-tap not supported in Tear?
2) On the long-press, my code is working as expected, but I'm also getting Tear's own page popop menu. Is there a way to suppress that popup?

Bundyo | # 1350 | 2011-06-13, 05:11 | Report

1) The dblclick event support is entirely up to the GtkWebKit engine to suport.
2) I'm not sure that contextmenu event is supported in that old version, and for sure the menu can't be canceled even if it is (I'm not sure it can in the other browsers available for NIT)

Check the PPK mobile events table here. You can see that dblclick event is not supported at all (check the legacy events part) and contextmenu event is barely supported across all mobile platforms. Knowing that all legacy events are simulated (that is - fired after the touch events, just for compatibility), this sounds rather logical.